At Web Summit Vancouver 2026, The Solutioners Labs Argues for AI That Actually Belongs Inside the Building

Vancouver on a Tuesday in May is grey-blue and a little cool, the kind of West Coast morning that softens the edges of everything. The Convention Centre sits right on the harbour. Seaplanes lift off behind the speakers’ tent. The mountains on the North Shore look closer than they are.

Twenty thousand people are walking past a giant white-and-blue "websummit Vancouver" wordmark on the plaza. They have come from a hundred countries for the second edition of Web Summit Vancouver — year two of a three-year run the city has committed to. Among them are Jawad Khalil and Nasir Teherany, the two founders of The Solutioners Labs, a Toronto-based firm that has spent the last eighteen months building what they call sovereign active intelligence for regulated industries.

They are not here to demo a new chatbot. They are here, by their own description, to argue against the chatbot — against the idea that AI is just a window you type into.

Ten Years Of Watching The Real Workflow

Before The Solutioners Labs, there was a cybersecurity practice. For more than a decade Jawad and Nasir’s team worked alongside law firms, banks, dealers, and accounting practices — the kinds of organizations where trust isn’t a marketing line, it is the entire business.

That decade taught them something most technology vendors never get to see: the real workflow. Not the workflow on the slide deck. The workflow at 4:50 p.m. on a Friday, when the deal needs to close and the partner is on a plane. The little gaps between systems. The compliance pressure sitting quietly behind every decision.

"You can’t architect for regulated work from the outside," Jawad says, standing near a coffee cart at the back of the AI Summit track. "You can wrap it. You can bolt onto it. But you can’t build for it until you’ve lived inside it long enough to understand what people actually do when the clock is ticking."

Then Commercial AI Arrived

Almost overnight, professionals at the firms they had spent a decade defending began pulling consumer AI tools into their day. Drafting client memos. Summarizing privileged correspondence. Pasting financial data into chat windows that were never designed for any of it.

From the outside, it looked like productivity. From where Jawad and Nasir stood — inside the compliance perimeters of those firms — it looked like a privacy problem and an ethics problem waiting to surface.

"The professionals weren’t being reckless," Nasir says. "They were being efficient. People do this with every new tool. The issue isn’t the user. The issue is that the tool, as shipped, doesn’t belong inside the building."

That observation is the founding story.

Active, Not Artificial

The Solutioners Labs was built in response. The thesis fits on one line: convert artificial intelligence into active intelligence.

"Models that sit in a chat window are artificial," Jawad says. "Systems that understand your context, run inside your compliance perimeter, and do real work for you — that’s active."

The difference is more than a slogan. An artificial-intelligence chatbot is something you have to remember to use, paste things into, and check after the fact. An active-intelligence system runs in the background — watching the case file, the policy folder, the advisor’s book — and surfaces what matters without being asked. Crucially, it does that without the data ever leaving the firm’s regulatory envelope. No outbound API call to a model the firm doesn’t control. No quiet logging of privileged content. No off-perimeter inference.

That last part is non-negotiable for the customers they serve. A Canadian advisor cannot ship a client portfolio to a US-hosted chat assistant and expect to remain on the right side of provincial regulation. A litigation partner cannot drop a draft pleading into a black-box completion endpoint and still call the result confidential. The whole point of the architecture is that they don’t have to.

A Conference Catching Up

The room they are walking through is, in a way, catching up to them. The big sessions on Vancouver’s schedule this week put the word sovereignty on the marquee without irony. Canada’s first Minister of Artificial Intelligence and Digital Innovation, Evan Solomon, takes Centre Stage with Cohere’s Joelle Pineau for a panel titled "AI for All: Canada’s AI Moment." B.C. Premier David Eby joins a separate panel on growing the province’s tech sector. The week’s biggest geopolitical announcement is the Canada–Germany Digital Alliance, signed in the building between sessions: a joint commitment on AI, digital sovereignty, infrastructure, and quantum.

The opening keynotes frame the whole conference as a single question: who owns the future? Who owns the models, the compute, the training data, the inference path. Three years ago that question was niche. This year it is the headline.

"We have spent two years arguing about model placement," Nasir says, watching the crowd thin between sessions. "Where the model lives. Whose perimeter it sits in. Whose data it can see. It is a strange feeling to look around the floor this week and find that everyone is finally having the same argument."

What That Looks Like In Practice

What The Solutioners Labs is actually building reads less like an AI product and more like back-office plumbing with a model threaded through it. A connector layer talks to the systems the firm already pays for — the back-office, the document management system, the CRM, the email archive. A sovereign model runs on infrastructure the firm itself can audit. An orchestration layer turns the model’s reasoning into concrete actions: draft this memo, surface this anomaly, flag this exception, suggest this rebalancing.

None of it is glamorous. None of it photographs well on a conference stage. That, Jawad and Nasir argue, is the point.

"Conference AI is built to demo," Jawad says. "Regulated AI is built to disappear into the work. The day a partner stops noticing the system is the day it has earned its keep."

The Quiet Bet

The real bet The Solutioners Labs is making is on the unfashionable kind of patience. The bet is that the firms who quietly built the boring infrastructure underneath the wow demos will be the ones still standing when the regulators show up. That the dealer who needs to scale tens of thousands of active client accounts across three regulatory codes will pick the platform that never requires a single confidential document to leave the building. That the law partner trying to keep three jurisdictions of privilege intact will choose the system whose architects spent a decade understanding what privilege actually is.

In Vancouver this week, that bet does not look as contrarian as it did a year ago. The hallway conversations have started to circle the same hard questions. The decks have started to use the same words.

"We don’t need to be the loudest company at the conference," Nasir says, as the harbour breeze picks up off the seawall. "We need to be the company that’s still standing when the regulators show up. That is a different optimization."

The clouds are moving. The mountains are sharper now. The second day of the second edition of Web Summit Vancouver is starting.